Stochastic non-stationary characteristics of vehicle-induced bridge vibrations

Abstract

Quantifying vehicle-bridge interaction mechanism is critical for bridge safety design, transportation weight control, condition assessments in maintenance, and structural health monitoring. In this paper, the stochastic non-stationary characteristics of a bridge under vehicle-bridge interaction are investigated. The response of the bridge under the loading of vehicles is, respectively, depicted in Fourier spectrum, spectrogram, and Hilbert spectrum. Both Fourier spectrum and spectrogram exhibit some cluster bands. The Hilbert spectrum exhibits time-varying frequencies in the proximity of low bridge natural modes, which suggests possible non-linear and non-stationary characteristics. The stochastic and non-stationary properties are further characterized by using bispectrum and recurrence analysis, and the weak nonlinear properties are characterized by using nonlinear invariants. The free-decay response of bridge (vehicle left the bridge) exhibits linear properties, which is correlated with the natural modes of bridge, whereas the bridge vibrations under moving vehicle exhibit complicated stochastic non-stationary and non-linear properties. These provide the insights into the conventional modal analysis-based approaches for heavy vehicle-loaded bridge condition assessments and structural health monitoring.